- 15 Jun, 2019 12 commits
-
-
David S. Miller authored
Eric Dumazet says: ==================== net/packet: better behavior under DDOS Using tcpdump (or other af_packet user) on a busy host can lead to catastrophic consequences, because suddenly, potentially all cpus are spinning on a contended spinlock. Both packet_rcv() and tpacket_rcv() grab the spinlock to eventually find there is no room for an additional packet. This patch series align packet_rcv() and tpacket_rcv() to both check if the queue is full before grabbing the spinlock. If the queue is full, they both increment a new atomic counter placed on a separate cache line to let readers drain the queue faster. There is still false sharing on this new atomic counter, we might in the future make it per cpu if there is interest. ==================== Acked-by:
Willem de Bruijn <willemb@google.com> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
Eric Dumazet authored
There are two places where we want to clear the pressure if possible, add a helper to make it more obvious. Signed-off-by:
Eric Dumazet <edumazet@google.com> Suggested-by:
Vinicius Costa Gomes <vinicius.gomes@intel.com> Signed-off-by:
-
Eric Dumazet authored
__packet_rcv_has_room() can now be run without lock being held. po->pressure is only a non persistent hint, we can mark all read/write accesses with READ_ONCE()/WRITE_ONCE() to document the fact that the field could be written without any synchronization. Signed-off-by:
-
Eric Dumazet authored
tpacket_rcv() can be hit under DDOS quite hard, since it will always grab a socket spinlock, to eventually find there is no room for an additional packet. Using tcpdump [1] on a busy host can lead to catastrophic consequences, because of all cpus spinning on a contended spinlock. This replicates a similar strategy used in packet_rcv() [1] Also some applications mistakenly use af_packet socket bound to ETH_P_ALL only to send packets. Receive queue is never drained and immediately full. Signed-off-by:
-
Eric Dumazet authored
Under DDOS, we want to be able to increment tp_drops without touching the spinlock. This will help readers to drain the receive queue slightly faster :/ Signed-off-by:
-
Eric Dumazet authored
Goal is use the helper without lock being held. Signed-off-by:
-
Eric Dumazet authored
Goal is to be able to use __tpacket_v3_has_room() without holding a lock. Signed-off-by:
-
Eric Dumazet authored
Goal is to be able to use __tpacket_has_room() without holding a lock. Signed-off-by:
-
Eric Dumazet authored
struct packet_sock is only read. Signed-off-by:
-
Robert Hancock authored
Commit "net: phy: Add detection of 1000BaseX link mode support" added support for not filtering out 1000BaseX mode from the PHY's supported modes in genphy_config_init, but we have to make a similar change in genphy_read_abilities in order to actually detect it as a supported mode in the first place. Add this in. Signed-off-by:
Robert Hancock <hancock@sedsystems.ca> Reviewed-by:
Florian Fainelli <f.fainelli@gmail.com> Signed-off-by:
-
Ivan Khoronzhuk authored
Only for consistency reasons, do it like in main cpsw.c module and use ndev reference but not by means of slave. Signed-off-by:
Ivan Khoronzhuk <ivan.khoronzhuk@linaro.org> Signed-off-by:
-
Ivan Khoronzhuk authored
No need to set ndev for drvdata when mainly cpsw reference is needed, so correct this legacy decision. Reviewed-by:
Grygorii Strashko <grygorii.strashko@ti.com> Signed-off-by:
-
- 14 Jun, 2019 28 commits
-
-
David S. Miller authored
Paolo Abeni says: ==================== net/mlx5: use indirect call wrappers The mlx5_core driver uses several indirect calls in fast-path, some of them are invoked on each ingress packet, even for the XDP-only traffic. This series leverage the indirect call wrappers infrastructure the avoid the expansive RETPOLINE overhead for 2 indirect calls in fast-path. Each call is addressed on a different patch, plus we need to introduce a couple of additional helpers to cope with the higher number of possible direct-call alternatives. v2 -> v3: - do not add more INDIRECT_CALL_* macros - use only the direct calls always available regardless of the mlx5 build options in the last patch v1 -> v2: - update the direct call list and use a macro to define it, as per Saeed suggestion. An intermediated additional macro is needed to allow arg list expansion - patch 2/3 is unchanged, as the generated code looks better this way than with possible alternative (dropping BP hits) ==================== Signed-off-by: -
Paolo Abeni authored
We can avoid another indirect call per packet wrapping the rx handler call with the proper helper. To ensure that even the last listed direct call experience measurable gain, despite the additional conditionals we must traverse before reaching it, I tested reversing the order of the listed options, with performance differences below noise level. Together with the previous indirect call patch, this gives ~6% performance improvement in raw UDP tput. v2 -> v3: - use only the direct calls always available regardless of the mlx5 build options - drop the direct call list macro, to keep the code as simple as possible for future rework v1 -> v2: - update the direct call list and use a macro to define it, as per Saeed suggestion. An intermediated additional macro is needed to allow arg list expansion Signed-off-by:
Paolo Abeni <pabeni@redhat.com> Acked-by:
Saeed Mahameed <saeedm@mellanox.com> Signed-off-by:
-
Paolo Abeni authored
We can avoid an indirect call per packet wrapping the skb creation with the appropriate helper. Signed-off-by:
-
David S. Miller authored
Merge tag 'mac80211-next-for-davem-2019-06-14' of git://git.kernel.org/pub/scm/linux/kernel/git/jberg/mac80211-next Johannes Berg says: ==================== Many changes all over: * HE (802.11ax) work continues * WPA3 offloads * work on extended key ID handling continues * fixes to honour AP supported rates with auth/assoc frames * nl80211 netlink policy improvements to fix some issues with strict validation on new commands with old attrs ==================== Signed-off-by:
-
Kevin Darbyshire-Bryant authored
Use extack error reporting mechanism in addition to returning -EINVAL NL_SET_ERR_* code shamelessy copy/paste/adjusted from act_pedit & sch_cake and used as reference as to what I should have done in the first place. Signed-off-by:
Kevin Darbyshire-Bryant <ldir@darbyshire-bryant.me.uk> Signed-off-by:
-
Greg Kroah-Hartman authored
When calling debugfs functions, there is no need to ever check the return value. The function can work or not, but the code logic should never do something different based on this. Also, there is no need to store the individual debugfs file name, just remove the whole directory all at once, saving a local variable. Cc: "David S. Miller" <davem@davemloft.net> Cc: Guillaume Nault <g.nault@alphalink.fr> Cc: netdev@vger.kernel.org Signed-off-by:
Greg Kroah-Hartman <gregkh@linuxfoundation.org> Acked-by:
Guillaume Nault <gnault@redhat.com> Signed-off-by:
-
David S. Miller authored
Heiner Kallweit says: ==================== r8169: add and use helper rtl_is_8168evl_up Few registers have been added or changed its purpose with version RTL8168e-vl, so create a helper for identifying chip versions from RTL8168e-vl. ==================== Signed-off-by: -
Heiner Kallweit authored
>From RTL8168e-vl the value in register MaxTxPacketSize is interpreted differently, therefore use new helper rtl_is_8168evl_up to set this register. Signed-off-by:
Heiner Kallweit <hkallweit1@gmail.com> Signed-off-by:
-
Heiner Kallweit authored
Add helper rtl_is_8168evl_up to make the code better readable and to simplify it. Signed-off-by:
-
James Prestwood authored
When the offchannel TX wait time expires, send the appropriate event. Signed-off-by:
James Prestwood <james.prestwood@linux.intel.com> Signed-off-by:
Johannes Berg <johannes.berg@intel.com>
-
James Prestwood authored
cfg80211_remain_on_channel_expired is used to notify userspace when the remain on channel duration expired by sending an event. There is no such equivalent to CMD_FRAME, where if offchannel and a duration is provided, the card will go offchannel for that duration. Currently there is no way for userspace to tell when that duration expired apart from setting an independent timeout. This timeout is quite erroneous as the kernel may not immediately send out the frame because of scheduling or work queue delays. In testing, it was found this timeout had to be quite large to accomidate any potential delays. A better solution is to have the kernel send an event when this duration has expired. There is already NL80211_CMD_FRAME_WAIT_CANCEL which can be used to cancel a NL80211_CMD_FRAME offchannel. Using this command matches perfectly to how NL80211_CMD_CANCEL_REMAIN_ON_CHANNEL works, where its both used to cancel and notify if the duration has expired. Signed-off-by:
-
Greg Kroah-Hartman authored
When calling debugfs functions, there is no need to ever check the return value. The function can work or not, but the code logic should never do something different based on this. Cc: Johannes Berg <johannes@sipsolutions.net> Cc: "David S. Miller" <davem@davemloft.net> Cc: linux-wireless@vger.kernel.org Cc: netdev@vger.kernel.org Signed-off-by:
-
Johannes Berg authored
This appears to happen occasionally, and if it does we really want even more information than we have now. Signed-off-by:
Luca Coelho <luciano.coelho@intel.com> Signed-off-by:
-
Johannes Berg authored
If HW advertises it has rate control, we skip all of the rate control assignments, but sometimes the data we have here is useful, especially so that we don't have to do the lookups again on which rates are configured and are supported. So do the low rate assignment anyway to help out drivers that might need it. Signed-off-by:
-
Johannes Berg authored
Even if we have a station, we currently call rate_control_send_low() with the NULL station unless further rate control (driver, minstrel) has been initialized. Change this so we can use more information about the station to use a better rate. For example, when we associate with an AP, we will now use the lowest rate it advertised as supported (that we can) rather than the lowest mandatory rate. This aligns our behaviour with most other 802.11 implementations. To make this possible, we need to also ensure that we have non-zero rates at all times, so in case we really have *nothing* pre-fill the supp_rates bitmap with the very lowest mandatory bitmap (11b and 11a on 2.4 and 5 GHz respectively). Additionally, hostapd appears to be giving us an empty supported rates bitmap (it can and should do better, since the STA must have supported for at least the basic rates in the BSS), so ignore any such bitmaps that would actually zero out the supp_rates, and in that case just keep the pre-filled mandatory rates. Signed-off-by:
-
Johannes Berg authored
There's no rate control algorithm that *doesn't* want to call it internally, and calling it internally will let us modify its behaviour in the future. Signed-off-by: -
Ilan Peer authored
Add the "OBSS Narrow Bandwidth RU In OFDMA Tolerance Support" flag definition to the definitions of the flags covered by the Extended Capability IE. Signed-off-by:
Ilan Peer <ilan.peer@intel.com> Signed-off-by:
-
Ilan Peer authored
Add a function that iterates over the BSS entries associated with a given wiphy and calls a callback for each iterated BSS. This can be used by drivers in various ways, e.g., to evaluate some property for all the BSSs in the medium. Signed-off-by:
-
John Crispin authored
Allow the userland daemon to en/disable TWT support for an AP. Signed-off-by:
Shashidhar Lakkavalli <slakkavalli@datto.com> Signed-off-by:
John Crispin <john@phrozen.org> [simplify parsing code] Signed-off-by:
-
John Crispin authored
Turn TWT for STA interfaces when they associate and/or receive a beacon where the twt_responder bit has changed. Signed-off-by:
-
Johannes Berg authored
Require that each vendor command give a policy of its sub-attributes in NL80211_ATTR_VENDOR_DATA, and then (stricly) check the contents, including the NLA_F_NESTED flag that we couldn't check on the outer layer because there we don't know yet. It is possible to use VENDOR_CMD_RAW_DATA for raw data, but then no nested data can be given (NLA_F_NESTED flag must be clear) and the data is just passed as is to the command. Signed-off-by: -
John Crispin authored
This function is similar to ieee80211_get_he_sta_cap() but allows passing the iftype. Also make ieee80211_get_he_sta_cap() use the new helper rather than duplicating the code. Signed-off-by:
-
Chung-Hsien Hsu authored
Let drivers advertise support for station-mode SAE authentication offload with a new NL80211_EXT_FEATURE_SAE_OFFLOAD flag. Signed-off-by:
Chung-Hsien Hsu <stanley.hsu@cypress.com> Signed-off-by:
Chi-Hsien Lin <chi-hsien.lin@cypress.com> Signed-off-by:
-
Chung-Hsien Hsu authored
Add definition of WPA version 3 for SAE authentication. Signed-off-by:
-
Chung-Hsien Hsu authored
Add NL80211_ATTR_IFINDEX attribute to port authorized event to indicate the operating interface of the device. Also put NL80211_ATTR_WIPHY attribute in it to be consistent with the other MLME notifications. Signed-off-by:
-
Alexander Wetzel authored
IEEE 802.11 - 2016 forbids mixing MPDUs with different keyIDs in one A-MPDU. Drivers supporting A-MPDUs and Extended Key ID must actively enforce that requirement due to the available two unicast keyIDs. Allow driver to signal mac80211 that they will not check the keyID in MPDUs when aggregating them and that they expect mac80211 to stop Tx aggregation when rekeying a connection using Extended Key ID. Signed-off-by:
Alexander Wetzel <alexander@wetzel-home.de> Signed-off-by:
-
Heiner Kallweit authored
tp->coalesce_info is used in rtl_coalesce_info() only, so we can remove this member. In addition replace phy_ethtool_get_link_ksettings with a direct access to tp->phydev->speed. Signed-off-by:
-
Heiner Kallweit authored
In case of a timeout currently ~0 is returned. Callers often just check whether a certain bit is set and therefore may behave incorrectly. So let's return -ETIMEDOUT in case of a timeout. r8168_phy_ocp_read is used in r8168g_mdio_read only, therefore we can apply the same change. Signed-off-by:
-
